1. Field of the Invention
The present invention relates to a technique of generating a fingerprint image to be used in fingerprint collation processing. In particular, the present invention relates to a technique of using a sensor mechanism to read a series of partial images representing a fingerprint from a finger which conducts slide operation and constructing a general image of the fingerprint from the series of partial images.
2. Description of the Related Art
As a kind of a conventional sensor mechanism to read a fingerprint image, there is the so-called sweep type fingerprint sensor. This sweep type fingerprint sensor has a sensor plane that is substantially the same in lateral width as a finger and that is shorter in longitudinal length than the whole fingerprint. The sweep type fingerprint sensor is a mechanism which continuously acquires partial images of a fingerprint from a finger which moves in substantially one direction. An image which can be acquired at each read timing of the sweep type fingerprint sensor has such a size as to be a part of the whole fingerprint. In order to acquire an image of the whole fingerprint, therefore, it is necessary to construct a general image joining a plurality of partial images successively read.
In constructing a general image from partial images of a fingerprint, there is, for example, a method of gradually shifting two consecutive partial images to overlap each other and determining a superposing position of both images on the basis of a difference between both images in the overlapping portion. In this method, a method of restricting a calculation subject of the difference to an inspection region which is a part of a partial image is adopted to fast calculate the difference between both images.
As a conventional method for determining an image region such as the inspection region, there is, for example, a method described in Japanese Patent Application Laid-Open No. 2003-208620. According to this method, a predetermined region in an input image, expected to overlap a preceding image is designated as a noted region, and a feature region corresponding to the inspection region is detected from the noted region. And a similar region that is the most similar to the feature region is detected from the preceding image, and the both images are combined so as to make the detected similar region and the feature region in the input image coincide with each other.
In the method described in Japanese Patent Application Laid-Open No. 2003-208620, however, the noted region, i.e., the region in the input image expected to overlap the preceding image is set in the same position for any input image. Only when the finger has slid in a predetermined one direction, therefore, it becomes possible to superpose the input image properly. For example, if the finger is slid in a direction opposite to the prescribed direction, therefore, a situation where the noted region in the input image does not correspond to the preceding image occurs and consequently the input image cannot be superposed on the preceding image. Furthermore, for example, when the finger does not slide smoothly on the sensor in one direction, a momentary motion of the finger substantially in the reverse direction is in fact caused. In such a case as well, however, a situation similar to the situation described above occurs. Therefore, it becomes difficult to construct a proper image of the whole fingerprint.